TMaxxx/agent-task-seta-env-v1.1

SETA-Env for tmax

Images require building: the dataset conversion and complete task archive are published. 0 task image(s) were published during preparation; the remaining images are intentionally left for the included resumable build script. No build job is running for this release. Only tasks with available images can run immediately. See image-build-status.json for the publication-time snapshot.

All 4,567 task directories from camel-ai/SETA-Env, pinned to 3c3bc8975b05826bf41769bd6b2da76ad1b6dd42, converted to tmax's swerl_vanillux_sandbox format.

The train split uses the same messages, ground_truth, dataset, env_config, and source schema as the other TMaxxx/agent-task-* datasets. Messages use the tmax Vanillux templates; dataset is passthrough. Task IDs are prefixed with seta_env__ to avoid collisions when combining task archives. task-manifest.json maps them back to original paths.

task-data.tar.gz contains a directory per task with all original task files preserved byte-for-byte, plus image.txt. Instructions, verifiers, build contexts, metadata, and reference solutions are retained. The tmax loader exposes only environment seed files and defers tests until submission; reference solutions are not mounted into the agent sandbox.

Usage

from datasets import load_dataset
from huggingface_hub import hf_hub_download

train = load_dataset("TMaxxx/agent-task-seta-env-v1.1", split="train")
archive = hf_hub_download("TMaxxx/agent-task-seta-env-v1.1", "task-data.tar.gz", repo_type="dataset")

Extract the archive and set the sandbox task_data_dir to the directory containing the task directories. Every row specifies its image explicitly. Task resource requirements and verifier timeouts remain in the original task.toml; configure the training harness accordingly.

Images

Images use hamishi740/agent-task-seta-env:<source-environment-hash> and target Linux AMD64. See task-manifest.json for build status.

Validation and attribution

All rows passed schema and Parquet round-trip checks, and every archived task file was checked against the downloaded source bytes. See validation.json. This does not constitute an evaluation of every task or a full training run.

tmax Vanillux prompts are adapted from mini-swe-agent (MIT).

Resume image builds

Requires Docker Buildx and a Docker Hub login with write access to the hamishi740 target repositories. The script skips tags already published, retains per-image logs, and can be rerun after failures.

mkdir -p tasks
tar -xzf task-data.tar.gz -C tasks
python build_dataset_images.py --task-data-dir tasks --manifest task-manifest.json --workers 8 --timeout 1800

To use another Docker Hub namespace, update the image references consistently in the manifest, training records, and task image.txt files before building.

Runtime smoke tests were not run for SETA because image builds were deferred. Structural, schema, and archive checks passed for all tasks.

Apptainer images

This release does not provide a hosted Apptainer image pool. Build the Docker images with the included build_dataset_images.py script above, then use apptainer pull docker://<image-reference> to create local SIFs. See task-manifest.json for the per-task image references and publication-time build status.

Source and license

This is a tmax training-format conversion of the original dataset. The original authors' license and reuse terms apply to the source data; consult the upstream license and dataset documentation for those terms, attribution, and citation requirements. This conversion does not relicense the source tasks or bundled third-party materials.